The Thermal Biology of Takydromus kuehnei Indicates Tropical Lizards From High Elevation Have Not Been Severely Threatened by Climate Change

Climate change poses different threats to animals across latitudes. Tropical species have been proposed to be more vulnerable to climate change. However, the responses of animals from tropical mountains to thermal variation and climate change have been scarcely studied. Here, we investigated the thermal biology traits of a tropical lizard (Takydromus kuehnei) distributed at high elevations (>950 m) and evaluated the vulnerabilities of T. kuehnei by thermal biology traits, thermal safety margin, and thermoregulatory effectiveness. The average active body temperatures of T. kuehnei in the field were 26.28°C and 30.65°C in April and June, respectively. The selected body temperature was 33.23°C, and the optimal temperature for locomotion was 30.60°C. The critical thermal minimum and critical thermal maximum temperatures were 4.79°C and 43.37°C, respectively. Accordingly, the thermal safety margin (1.23°C) and thermoregulatory effectiveness (1.23°C) predicted that T. kuehnei distributed in tropical mountains were not significantly depressed by environmental temperatures. This study implies that high-elevation species in tropical regions may not be severely threatened by ongoing climate change and highlights the importance of thermal biology traits in evaluating the vulnerability of species to climate change.


INTRODUCTION
Climate change has negatively affected animal distribution and abundance (Root et al., 2003;Thomas et al., 2004;Medina et al., 2016). Although the latitudinal pattern of the vulnerabilities of animals to climate change is still controversial, increasing investigations claim that tropical animals are profoundly vulnerable to climate change, considering higher metabolic rates (e.g., Dillon et al., 2010), narrow thermal-safety margins (TSM) (e.g., Sunday et al., 2014), and depressed life-history cycles (e.g., Blouin-Demers and Weatherhead, 2001;Deutsch et al., 2008). For instance, the environmental temperature in tropical regions is higher than that in other latitudes, making it possible to exceed the physiological thermal-tolerance limits of animals (Huey et al., 2009;Sunday et al., 2014).
However, animals can migrate toward high latitudes and high elevations to avoid the risk of being threatened by climate change (e.g., Forero-Medina et al., 2011;Freeman et al., 2018). High latitudes and elevations can provide lower average temperatures and more fluctuating temperatures, which provide retreat (i.e., cool places) for thermoregulation. Therefore, tropical mountains may provide refuge for tropical animals to escape from exposure to warming temperatures (Bonebrake and Deutsch, 2012). However, it is largely unknown whether the highelevation species in tropical regions are depressed by ongoing climate change (Freeman et al., 2018). Understanding the vulnerabilities of animals from tropical mountains is important in not only revealing the thermal adaptation of tropical species at high elevations but also evaluating the availabilities of migrating toward high elevations to escape from climate change (Ghalambor et al., 2006).
As ectotherms, reptiles have been threatened by climate change in the past decades and are projected to encounter more severe threats in the future Sinervo et al., 2010;Barnosky et al., 2011;Diele-Viegas et al., 2020;Taylor et al., 2020). Lizards regulate body temperatures by an external heat source, and thus, their biological functions are subject to the effect of thermal environments on body temperatures (Huey, 1982;Hertz et al., 1993). Therefore, compared to other taxa, lizards are particularly sensitive to thermal variations and constitute an appropriate study system for evaluating the vulnerability of animals to climate change (Huey et al., 2012;Taylor et al., 2020). The thermal biology traits, thermal safety margin, and thermoregulation effectiveness are important for thermal adaptation and are also indispensable proxies for determining the vulnerability to climate change in lizards (Huey et al., 2012;Clusella-Trullas et al., 2021). The effects of climate change on lizards depend on the interaction of thermal environments and the integration of thermal biology traits, thermoregulation effectiveness in maintaining body temperature, and thermal safety margins (Sunday et al., 2014;ObregÓn et al., 2020). For example, lizards would maintain their active body temperature in an appropriate range by thermoregulation, which is approximate to the range of optimal body temperatures for various physiological processes (Huey and Kingsolver, 1989;Angilletta et al., 2002). The critical thermal minimum (CT min ) and critical thermal maximum (CT max ) temperatures are thermal limits of permitting performance (Angilletta et al., 2002;Taylor et al., 2020), which generally depend on the thermal environment of species (Ji et al., 1996;Chen et al., 2003;Zhang and Ji, 2004). In addition, the thermal sensitivity of locomotion and resting metabolic rates (RMR) is essential for evaluating the thermal adaptation in ectotherms (e.g., Dillon et al., 2010;Shu et al., 2010;Sun et al., 2014Sun et al., , 2018b. Therefore, thermal biology traits, including body temperature, selected body temperature range, and thermal tolerance, can be integrated to determine the thermoregulatory effectiveness and TSM, which are critical for evaluating the vulnerability of reptiles to climate change (Hertz et al., 1993;Sunday et al., 2014).
The Takydromus lizards have a wide distribution across latitudes in China, ranging from tropical to cold-temperature regions (Zhao, 1993;Lin et al., 2002;Portniagina et al., 2019). Takydromus kuehnei is a small lacertid lizard (snout-vent length, SVL < 60 mm), mainly distributed in tropical regions across southern China. The population on Hainan Island is a typically tropical population in China. According to previous predictions of species vulnerability across latitudes, T. kuehnei of the Hainan population may be the most vulnerable in Takydromus lizards to climate change in China. In the Diaoluoshan Mountains, T. kuehnei is distributed across low to high elevations ranging from 400 to 1000 m (Wang, 2014). Therefore, the high-elevation population of T. kuehnei is an excellent study system for analyzing the response of high-elevation tropical species to thermal variation, considering the interaction between warming climates and thermal biology traits.
With high-elevation population of T. kuehnei from the Diaoluoshan Mountains, this study monitored thermal environments, determined thermal biology traits, and calculated thermoregulatory effectiveness and TSM. By comparing the thermal biology traits, thermoregulatory effectiveness, and TSM with published data across latitudes (e.g., Zhang and Ji, 2004;Sunday et al., 2014;Hao et al., 2020), we aimed to evaluate the vulnerabilities of tropical species (i.e., T. kuehnei in this study) from high elevation to climate change and thus test whether mountains are potential refuges for tropical species under ongoing climate change. Base on the assumption that the ambient temperatures at high elevation in tropical areas are low on average with high variations, which are similar to that from medium and high latitudes (e.g., Ghalambor et al., 2006;Freeman et al., 2018), and the effects of thermal environments on thermal biology traits in Takydromus lizards (Hao et al., 2020), we predicted that the thermal biology traits of T. kuehnei from high elevation are similar to congeners from medium or high latitudes. Accordingly, we also predicted that T. kuehnei are not vulnerable to climate change by high thermoregulation effectiveness and big thermal safety margin. Further, we predicted that the high elevations (i.e., tropical mountains) can be potential refuges for tropical species to escape from exposure to climate change.

Study Species Collection and Active Body Temperatures in the Field
Fieldwork was performed at the Mt. Diaoluo National Reserve in Hainan, China (18 • 43 N, 109 • 52 E). We collected 20 adult T. kuehnei (12 males and 8 females) from a high-elevation population (height > 900 m) using either a noose or by hand.
During collection, we measured the active body temperature of T. kuehnei once the lizard was captured. We measured cloacal temperatures using an electronic thermocouple within 30 s of capture (UNT T-325, Shanghai, China). The active body temperatures of the lizards were also measured repeatedly in June when they were released.

Operative Temperatures
During lizard collection, we also monitored the operative temperatures (T e ) using copper tube models. The copper tube models mimicked the size of the lizards (diameter × length: 15 mm × 70 mm). Each model was inserted with an iButton (DS1921, MAXIM Integrated Products Ltd., United States). We randomly set the copper models exposed to full sun and shaded sites. The iButtons recorded the hourly temperature. We used two iButtons for each site, and the average of two iButtons for each site was calculated as T e . The T e collection lasted from April to June when the lizards were released.

Lizard Husbandry
After collection in April, we transported the lizards to the laboratory built in the Mt. Diaoluo National Reserve. We weighed (body mass, ± 0.01 g) and measured (SVL, ± 0.1 mm) the lizards and arranged every five lizards with different sexes in each plastic terrarium (150 cm × 40 cm × 50 cm, length × width × height) with moist soil. The terraria were set in a room where the temperature was maintained at 18 ± 1 • C. A supplementary heating lamp was suspended above one end of each terrarium to create temperatures ranging from 18 to 40 • C from 06:00 to 20:00. Food (crickets and larval Tenebrio molitor) and water were provided ad libitum. After three days of rearing, we started to determine the thermal biology traits. During the test, if the females were pregnant, we supplemented the test of the traits after they laying the eggs.

Selected Body Temperature
The measurements of T sel were also conducted in a temperature-controlled room at 18 ± 1 • C. To create a thermal gradient, we placed a large tank (1000 × 500 × 300 mm, length × width × height) and set a 275 W incandescent lamp above one end of the tank. The heating period was 06:00 to 20:00, which mimicked the natural light period. We followed two established protocols to measure the selected body temperature (T sel ). To facilitate comparison of the T sel of T. kuehnei to published data from other Takydromus lizards, we first followed an old protocol (Ji et al., 1996;Zhang and Ji, 2004). Briefly, four to five lizards were introduced into the cold end of the terrarium at 16:00 on the first day. The next day, we measured body temperature (cloacae, within 30 s) for each individual at 09:00 and 15:00. The average of two measurements for each lizard was calculated as the selected body temperature (T sel ). To facilitate thermoregulatory accuracy and effectiveness comparison, we also determined the selected temperature range using a new protocol (Li et al., 2017). In brief, we measured body temperatures (cloacae, within 30 s) at each hour from 08:00 to 18:00. The selected temperature range was calculated as the central 50% of all recordings for each lizard.

Thermal Tolerance
We determined the critical thermal minimum (CT min ) and critical thermal maximum (CT max ) in a programmed incubator (KB 240,Binder,Germany). Before determination, the candidate was acclimated to 28 • C for 2 h. Thereafter, the lizards were cooled/heated from 28 • C at a rate of 1 • C per 10 min. During the cooling/heating period, we constantly monitored the righting responses of the lizards. Once lizards lost the capacity to right themselves after being turned over, the body temperature (cloacae) was recorded as CT min /CT max . Thereafter, the lizards were immediately moved to 28 • C for recovery. If lizards could not recover from the test, their records were eliminated from further analysis. In this study, all lizards recovered to normal after cooling or heating.

Locomotor Performance
The locomotor performance of lizards was estimated by sprint speed at six test temperatures ranging from 18 to 38 • C (18,22,26,30,34, and 38 • C, in a randomized sequence). Before the measurement, the lizards were acclimated in a temperaturecontrolled incubator at the test temperature for 2 h. Sprint speed was measured in a customer-made wood racetrack (1200 mm × 100 mm × 150 mm, length × width × height, with intervals marked every 200 mm) in a temperaturecontrolled room. The lizard was introduced into the racetrack from one end and stimulated by a paintbrush to run through the racetrack. The running processes were recorded using an HD video camera (DCR-SR220E, Sony, Japan). Each lizard was tested twice at the test temperature with an interval of 1 h for rest. We analyzed the video of locomotion with established methods (Shu et al., 2010;Sun et al., 2014). For each lizard, we recorded the fastest speed at 200 mm intervals for each test by raw numbers, and the average of the fastest records from each test was calculated as the sprint speed.

Resting Metabolic Rates
The resting metabolic rate of T. kuehnei was determined at six test temperatures (18,22,26,30,34, and 38 • C) in a random order using a respirometry system (Sable Systems International, Henderson, NV, United States). We estimated the RMR by measuring the CO 2 production rate as a proxy for published protocols (Sun et al., 2018b(Sun et al., , 2020. The lizards were fasted for at least 12 h before the test and were acclimated for 2 h at the test temperature. The RMR of the lizards was determined using a closed-circuit system (volume = 281.4 mL). The lizard was housed at the test temperature in the chamber, which was placed in a temperature-controlled incubator (MIR554-PC, Sanyo, Japan). First, we set the system open to the air that had sucked up moisture through a tube with 300 ml/min flow rate to stabilize the baseline. After 3 min of opening to the air, the circuit system was transferred to be closed. We continuously recorded the rate of carbon dioxide production (VCO 2 ) for at least 7 min in a closed-circuit system. The RMR was calculated as the carbon dioxide production per gram of body mass per hour (mL g −1 h −1 ), using the equation RMR = VCO 2 × volume/body mass, where VCO 2 is the CO 2 production rate in percentage (%/h) in the closed-circuit system.

Thermoregulatory Accuracy and Effectiveness, and Thermal Safety Margin
For the calculation of the thermoregulation accuracy (d b ), we followed published protocols (Li et al., 2017). If T b was below (or above) the selected temperature range, d b was calculated as the difference between the T b and the lower (or upper) bound of the selected temperature range. Alternatively, if T b was within the selected temperature range, d b was zero. The T e values at day time from 07:00 to 17:00 were used to calculate de. If T e was below (or above) the selected temperature range, d e was calculated as the difference between T e and the lower (or upper) bound of the selected temperature range. When T e was within the selected temperature range, d e was zero. Finally, we used two indices, E and d e -d b , to estimate thermoregulatory effectiveness. E was calculated using the equation E = 1mean d b/ mean d e , following published methods (Hertz et al., 1993;Li et al., 2017). The thermal safety margins (TSM) were calculated as the difference between T e,max (the mean maximum hourly T e ) and CT max , following the published protocol (Sunday et al., 2014).

Statistical Analysis
The normality and homogeneity were checked using the Shapiro-Wilk test and Bartlett's test, respectively. First, we analyzed sex differences in T sel, CT min , and CT max by one-way ANOVA. Thereafter, we analyzed the T a and T e with Kruskal-test between April and June, because they are independent in test time. We analyzed the difference in T e between the open and shaded locations using the Wilcoxon test. We tested between-month differences in T a , T e , d b , and d e using the Wilcoxon test. We also analyzed the difference between T sel and T a, and T sel and T e using the Wilcoxon test. For the analysis of locomotion, we used the modified Gaussian regression to fit the dependence of sprint speed on test temperatures using a published method (Hao et al., 2020). Subsequently, the optimal temperature for locomotion was calculated according to the regression. The thermal dependence of the resting metabolic rates was analyzed using the allometric growth curve to body temperature. If there was a significant difference between sexes in thermal biology traits, we supplementarily show the data with males and females separately.

Thermal Biology Traits
The selected body temperature (T sel ) for T. kuehnei was 33.23 ± 0.28 • C by the old protocol. The range of T sel for T. kuehnei was 28.2 • C to 36.4 • C, with the average value of T sel being 32.12 ± 0.21 • C by the new protocol. Sex did not affect T sel by the old (F 1 , 17 = 0.705, p = 0.413) or new protocol (F 1 , 68 = 0.495, p = 0.484) ( Table 1). The CT min and CT max for T. kuehnei were 4.79 ± 0.18 • C and 42.37 ± 0.11 • C, respectively. We found significant differences in CT min (F 1 , 17 = 8.828,  p = 0.009) and CT max (F 1 , 17 = 6.241, p = 0.023) between males and females ( Table 1).

Locomotor Performance and Metabolic Rates
The sprint speed for T. kuehnei was significantly dependent on body temperature; it increased with body temperature within the range of 18 to 30 • C and then decreased at higher body temperatures from 30 to 38 • C. According to the modified Gaussian regression, the optimal body temperature for maximum sprint speed was 30.60 • C ( Figure 3A). Metabolic rates were dependent on body temperature, with increasing allometrically against test temperatures (R 2 = 0.82, p < 0.0001) ( Figure 3B).

Thermoregulatory Accuracy, Effectiveness, and Thermal Safety Margin
Operative temperatures was significantly lower than the average T sel (W = 20659, p < 0.001). Most T e values (N = 706) were below the lower threshold of the selected temperature range, 156 values were higher than the upper threshold of the selected temperature range, and 380 values were within the range of selected body temperatures. Accordingly, d e did not vary between seasons (W = 189527, p = 0.8171). T a was lower than T sel in both April (W = 797, p < 0.001) and June (W = 2541, p < 0.001). In April, most T a values (N = 115) were below the selected temperature range, and the remaining values (N = 28) were within the selected temperature range. In June, most T a values (N = 109) were in the selected temperature range, and four values were higher than those above the selected temperature range. Accordingly, the d b in June was lower than in April (W = 15748, p < 0.001, Table 1). In addition, the d e of T. kuehnei was higher than that of the d b in April and June. Therefore, the d e -d b and E values were positive (Table 1). Similarly, all TSM values were positive, and the average TSM for T. kuehnei was 1.23 • C (Figure 4 and Table 1).

DISCUSSION
Tropical (i.e., low latitude) species are proposed to be more vulnerable to climate change than those from subtropical and temperate regions (i.e., medium and high latitudes) (e.g., Deutsch et al., 2008;Dillon et al., 2010;Sunday et al., 2019). However, high-elevation regions (i.e., tropical mountains) are potential refuges for species under climate change (Forero- Medina et al., 2011;Freeman et al., 2018;Meng et al., 2019). Therefore, it is critical to determine the thermal biology and thus the vulnerabilities to climate change of species from tropical mountains (i.e., high elevations), providing a reference for species migration toward high elevations (Huey et al., 2009). According to the thermal biology traits, thermoregulatory effectiveness, and thermal safety margin, we found that the high-elevation population of T. kuehnei is not a severe threat to climate change, implying that tropical mountains are available for tropical species in escaping the exposure to climate change (e.g., Freeman et al., 2018).
FIGURE 4 | Thermal-safety margins (TSM) of lizards across latitudes. Each spot indicates a TSM of each species and the red triangle represents the TSM of high-elevation T. kuehnei. The orange line represents the fitting curve of TSM along the global latitude and the gray area represents the 95% confidence interval. The database was collected from published analysis (Sunday et al., 2014).

Thermal Biology Traits of Takydromus kuehnei Did Not Follow the Existing Latitudinal Pattern
As ectotherms, most aspects of behavior and physiology in reptiles are profoundly affected by temperature, making reptiles an important model for studying the vulnerabilities of species to climate change (Huey et al., , 2012. Accordingly, numerous studies have focused on the response of reptiles to thermal variation and climate change (see details in Huey and Berrigan, 2001;Taylor et al., 2020). Comparing biological traits among populations and species is an essential way to evaluate the responses of species to climate change across geographical clues (Bacigalupe et al., 2018;Taylor et al., 2020). Takydromus lizards are distributed from the northernmost Heilongjiang province to the southernmost Hainan province in China Cai et al., 2012;Ma et al., 2019;Portniagina et al., 2019). Previous studies have summarized the thermal biology traits of the genus Takydromus distributed in China across latitudes (Hao et al., 2020). The T sel of the genus Takydromus lizards tends to decrease toward low latitudes. However, in contrast to the current pattern, we found that T sel of T. kuehnei from the high-elevation population is 33.23 ± 0.28 • C, which is the highest among the five Takydromus lizards with published data collected by the same protocols (i.e., old protocol in this study) (Ji et al., 1996;Chen et al., 2003;Zhang and Ji, 2004;Hao et al., 2020). Similarly, thermal tolerance does not fit the current latitudinal pattern. Our previous study indicated that the thermal tolerance range increased toward high latitudes by increasing CT max and decreasing CT min (Hao et al., 2020). However, in the high-elevation population of T. kuehnei, CT max and CT min are similar to those of medium-altitude species (i.e., T. septentrionalis) (Ji et al., 1996;Hao et al., 2020).

Thermoregulatory Effectiveness and Thermal Safety Margin Indicate High-Elevation Takydromus kuehnei Are Not Seriously Threatened by Climate Change
In this study, the T a for T. kuehnei from the high-elevation population was lower than congeners (i.e., T. septentrionalis, Sun et al., 2018a) in both April (26.28 • C) and June (30.65 • C), although it is a tropical lizard under high T e (Figures 2A,B).
In addition, although T a exhibited significant seasonal variation, only 1.6% (4/246) records were higher than the selected body temperature range, which is opposite to the current situation in tropical species where most body temperatures are near or higher than selected body temperatures (Huey et al., 2009). Furthermore, d b in June was lower than in April, indicating that the lizards in June expressed higher thermoregulatory accuracy in maintaining their body temperatures within the range of thermal preference. The seasonal variation in higher T a under lower T e (Table 1 and Figures 1, 2) appears opposite to the general knowledge that the body temperatures of lizards are subject to ambient temperatures (Huey, 1982). Lizards have different thermoregulatory strategies to adapt to seasonal temperatures (Vicente Liz et al., 2019). Lizards may select a low body temperature in the hot season to avoid exposure to ambient heat temperatures (Firth and Belan, 1998). The higher selected body temperatures in the mild season are probably because mild-season temperatures can relieve lizard exposure to the critical thermal limits (Huey, 1982;Vicente Liz et al., 2019). Accordingly, lizards may enhance activity and facilitate the maintenance of the optimal temperature at a low cost in a mild season (Huey and Slatkin, 1976). In this study, the thermoregulatory effectiveness (i.e., E) of high-elevation T. kuehnei in June was higher than that in April ( Thermal-safety margins has been considered as a primary index for evaluating the vulnerability of species to extreme heat (e.g., Deutsch et al., 2008;Sunday et al., 2014). Previous studies revealed that TSM decreased toward low latitudes (i.e., tropical areas), where the TSM of most species was below zero (see details in Sunday et al., 2014). However, this study found that the TSM of high-elevation T. kuehnei is higher than zero, as high as that of lizards from medium and high latitudes around 40 • (Figure 4). The TSM was determined using CT max and T e max. However, the CT max in lizards is conservative across latitudes because lizards cannot effectively increase the temperature tolerance in the face of rapidly rising temperatures due to climate change (e.g., Sunday et al., 2011;Logan et al., 2014). Therefore, the mild thermal environment, thus T e , can result in a high TSM for high-elevation T. kuehnei.
Generally, high thermoregulatory accuracy and large TSM predict high-elevation T. kuehnei would not be threatened by ongoing climate change in the future. Higher E allows T. kuehnei to regulate body temperatures with high effectiveness under thermal fluctuance (e.g., Li et al., 2017), and large TSM indicates that T. kuehnei is not under the stress of extreme heat (e.g., Sunday et al., 2014).

Tropical Mountains Are Plausible Refuges for Tropical Species
The primary reason for the lower risk of threat of T. kuehnei in this study may be that the population is collected from high elevations (>900m). In contrast, other Takydromus lizards were collected from low elevations (i.e., ∼200 m to ∼ 290 m) (Hao et al., 2020). High elevation in tropical regions can provide more fluctuant but lower average temperatures than low elevations (e.g., Leuschner, 2000;de Carvalho et al., 2019), which facilitates behavioral thermoregulation and increases the TSM. A large TSM at high elevation is possibly induced by decreased T e max, although normal CT max (Sunday et al., 2019). The climate variability hypothesis (CVH) posits that tropical organisms in warm and stable thermal environments should possess lower plasticity in their behavioral and physiological responses to thermal variation (Ghalambor et al., 2006;Gaston et al., 2009). In contrast, species from high-latitude or high-elevation habitats in more fluctuating thermal environments, and thus can tolerate or thrive over a broader range of temperatures, indicating a larger thermal safety margin due to being more resilient to climate change (Addobediako et al., 2000;Sunday et al., 2019). For example, the population of Podarcis hispanica from high elevation selected significantly higher body temperatures in the thermal gradient than the population with low elevation (Gabirot et al., 2013).
In summary, T. kuehnei from high elevation expresses similar thermal biology traits to medium-latitude congeners and allows high thermoregulatory accuracy and effectiveness, as well as a large thermal safety margin. Therefore, T. kuehnei could maintain a safe state in tropical regions under climate change. However, this study has some limitations. First, we only focused on the alpine population (>900m), lacking a comparison to the lowelevation population, or some other species with a narrow distribution in elevation/latitude. In the future, further studies with more tropical populations across elevations are needed to understand the vulnerability of tropical species to climate change (Ghalambor et al., 2006). With more intra-and interpopulation comparisons, we can also understand the diversity of adaptive strategies to climate change of tropical species (Huey et al., 2009).

DATA AVAILABILITY STATEMENT
The data supporting the conclusions of this article can be found in the Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by Animal Ethics Committees at the Institute of Zoology, CAS (IOZ14001).